empty field label constructor infelicity

[Hal Daume III]

Hi again,

The report says "The expression F {}, where F is a data constructor, is
legal whether or not F was declared with record syntax, provided F has no
strict fields: it denotes F _|_1 ... _|_n where n is the arity of F."

It unclear to me why there needs to be this provision for records with
strict fields -- just let them be undefined -- but that notwithstanding,
GHC seems to do the wrong thing:

> module Foo where
> data F = F !Int deriving (Show, Eq)
> data G = G  Int deriving (Show, Eq)

If we then load it up in ghci, both "F {}" and "G {}" ellicit the same
error: "Missing field in record construction".

If we do:

Foo> case (F {}) of { F x -> "1" }

we get the exeption.  If we use G instead of F, we correctly get "1".

Hugs seems to obey the report.  For "F {}" it
give: "INTERNAL_ERROR: depConFlds" which is different from "G {}" which
yields "G" followed by "Program error: {undefined}".

For the case expressions, Hugs generates "INTERNAL ERROR: depConFlds" for
F and "1" for G.

Arguably, this is weirdness in the report, but I think it's clear that GHC
isn't doing the right thing (where right thing is defined to be what the
report says).

 - Hal

--
Hal Daume III

 "Computer science is no more about computers    | hdaume@isi.edu
  than astronomy is about telescopes." -Dijkstra | www.isi.edu/~hdaume